Parkinson’s Disease Evaluation

Important

This report is intended to assist qualified Health Care Profesionals (HCP) in the assessment of an individual referred under the suspicion of having Parkinson’s Disease.

Clinical Context

This report presents several AI metrics derived from objective measures from individuals performing a battery of test using Manus Neurodynamica NeuroMotor Pen^TM.

These presented metrics combine factors from detailed measurement recordings made whilst the individual performs a battery of well established neurological test tasks.

The metrics have been assessed in a UK reference population and an individual’s results are presented in this clinical context. The HCP should review Clinical,Reference, Study et al [1] to establish applicability and limitations.

The information in this report should be used in the context of a full neurological assessment conducted to the current standard of care practices to establish a diagnosis.

Subject and Recording Details

Subject ID	NMC043
Sex	MALE
Dominant hand	LEFT
Test date and time	2022-08-11 16:56
Test battery	Circle, Spiral, Elel, FITTS Short Modified, FITTS Long Modified, ZigZag, ZigZag Offset
Operator comment	old pacient
	tbl Madopar 250mg 3x1/4

Overall Assessment

Subject performance similar to PD population (recommend review of report details)

Clinical context

In the clinical reference population, 20 individuals with a value less than 0.58 were subsequently diagnosed with PD. That is, 47.62% of the PD diagnoses in the study.

Additionally, 0 individuals with a value greater or equal to 0.58 were subsequently diagnosed as not having PD. That is, 0.0% of the non PD diagnoses in the study.

Note

We need to look into this whole ‘confidence’ thing. It is really confusing. In the above plot I have shown 1.0 - reported conf for the ‘predicted as non PD individuals’. I am not happy with the obvious 0.5 clipping above. We would expect this in a classifier operating on it’s own training data and not a real world clinical test.

Symptom Scores

These mini boxplots show the scores in a clinical context. Currently against the ‘Walker study’ data. A bigger pool would be much better (so max 83 individuals, usually lower if raw data did not result in successful classification).

Micrographia Features

The micrographia symptom assessment is derived from a combination of factors in the elel task.

{'FN': 26, 'TN': 22, 'TP': 16, 'FP': 8}

Tremor Features

{'FN': 3, 'TN': 4, 'TP': 39, 'FP': 26}

Bradykinesia

{'FN': 29, 'TN': 25, 'TP': 13, 'FP': 5}

Spatial Accuracy

{'FN': 15, 'TN': 9, 'TP': 27, 'FP': 21}

Test Battery Details

2023-03-09 16:51:51.784 | INFO     | neuromotor_pen.data:_parse_header:693 - Loading New Data Version...

ManusData_pd_NMC043_11082022052559.JSON

Circle

Circle Segment 1

Duration 24.6 s, Accuracy Estimate 0.903 (lower is better)

Circle Segment 2

Duration 18.36 s, Accuracy Estimate 0.779 (lower is better)

Circle Segment 3

Duration 23.56 s, Accuracy Estimate 1.154 (lower is better)

Circle Segment 4

Duration 19.0 s, Accuracy Estimate 0.827 (lower is better)

Circle Segment 5

Duration 21.8 s, Accuracy Estimate 1.096 (lower is better)

Circle Segment 6

Duration 20.4 s, Accuracy Estimate 1.092 (lower is better)

Circle Segment 7

Duration 15.04 s, Accuracy Estimate 1.224 (lower is better)

Circle Segment 8

Duration 18.32 s, Accuracy Estimate 1.014 (lower is better)

Circle Segment 9

Duration 18.96 s, Accuracy Estimate 1.157 (lower is better)

Circle Segment 10

Duration 15.8 s, Accuracy Estimate 2.256 (lower is better)

Spiral

Spiral Segment 1

Duration 49.2 s, Accuracy Estimate 1.45 (lower is better)

Spiral Segment 2

Duration 48.16 s, Accuracy Estimate 1.579 (lower is better)

Spiral Segment 3

Duration 52.16 s, Accuracy Estimate 1.767 (lower is better)

Spiral Segment 4

Duration 60.44 s, Accuracy Estimate 2.455 (lower is better)

Elel

Elel Segment 1

Elel Segment 2

Elel Segment 3

Elel Segment 4

Elel Segment 5

Elel Segment 6

Elel Segment 7

Elel Segment 8

Elel Segment 9

Elel Segment 10

Elel Segment 11

FITTS Short Modified

Can’t do fitts_short because low correlation between both touching-pressure and force-pressure

FITTS Long Modified

ZigZag

ZigZag Segment 1

Duration 32.52 s, Accuracy Estimate 3.128 (lower is better)

ZigZag Segment 2

Duration 42.04 s, Accuracy Estimate 2.67 (lower is better)

ZigZag Segment 3

Duration 37.12 s, Accuracy Estimate 2.809 (lower is better)

ZigZag Segment 4

Duration 33.76 s, Accuracy Estimate 2.527 (lower is better)

ZigZag Segment 5

Duration 34.08 s, Accuracy Estimate 3.034 (lower is better)

ZigZag Offset

ZigZag Offset Segment 1

Duration 23.04 s, Accuracy Estimate 0.067 (lower is better)

ZigZag Offset Segment 2

Duration 24.0 s, Accuracy Estimate 0.04 (lower is better)

ZigZag Offset Segment 3

Duration 18.12 s, Accuracy Estimate 0.03 (lower is better)

ZigZag Offset Segment 4

Duration 18.52 s, Accuracy Estimate 0.047 (lower is better)

ZigZag Offset Segment 5

Duration 17.56 s, Accuracy Estimate 0.058 (lower is better)

Appendices

Misc

Currently a dumping ground for things that could be included or previous output style.

Note

Putting all the results out here but will not be in a final report.

HiSpec	{‘HiSpec_class’: ‘NOT PD’, ‘HiSpec_score’: 0.73}
RanFor	{‘RanFor_class’: ‘PD’, ‘RanFor_score’: 0.58}
BM_May22	{‘BM_May22_class’: ‘NOT PD’, ‘BM_May22_score’: 0.1527203815031961}
BM_HC_Sep22	{‘BM_HC_Sep22_class’: ‘Patient’, ‘BM_HC_Sep22_score’: 0.8943770216925635}
BM_PD_Sep22	{‘BM_PD_Sep22_class’: ‘PD’, ‘BM_PD_Sep22_score’: 0.6254813892434572}